Turn Ghostwriter's flat message list into a tree structure, enabling
Claude.ai/ChatGPT-style branching conversations.
**Data model**: Add `parentMessageId` and `activeChildId` to messages,
`activeRootMessageId` to threads. Migration backfills existing messages
into a linear chain and links regenerated messages as siblings.
**Backend**: Tree-walking queries (getActivePathFromRoot, getAncestorPath,
getSiblingsOf), rewritten history builder that follows the ancestor path,
new editMessage and switchBranch services, regenerate now works on any
assistant message (not just the latest).
**Frontend**: BranchNavigator component (← 2/3 → arrows), inline edit on
user messages, per-message regenerate on assistant messages, regenerate
button removed from composer (now per-message).
**Infra**: Pin Node 22 via Volta to prevent ABI mismatches with
better-sqlite3 across environments.
When switching between job tabs (e.g. Discovered → Applied), the detail
panel would briefly show the previously selected job with the new tab's
action buttons — confusing and error-prone.
Three coordinated fixes:
- setActiveTab now checks if the selected job's status fits the target
tab, keeping it when valid (e.g. Discovered → All Jobs) and clearing
it otherwise.
- New visibleSelectedJob useMemo guard synchronously nulls out the
selected job when it doesn't belong to the active tab, eliminating
the one-frame flash caused by the data hook's useEffect lag.
- Auto-select effect now handles the case where a job passes the status
check but gets filtered out by source/sponsor/salary filters.
* Add startup.jobs extractor support
* Harden startup.jobs extractor inputs
* Wire startupjobs into Docker and CI
* Tighten startupjobs review follow-ups
* fix: publish ghcr during release workflow
* feat: add startupjobs max jobs configuration and update related tests
* Add RxResume URL setting to dashboard
* Potential fix for pull request finding
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
* Potential fix for pull request finding
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
* Potential fix for pull request finding
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
* Potential fix for pull request finding
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
---------
Co-authored-by: Copilot Autofix powered by AI <175728472+Copilot@users.noreply.github.com>
* Add language settings for AI-generated resume output
* Resolve merge conflicts for language settings PR
* Fix language settings review feedback and CI lint
* Tighten language setting precedence and onboarding validation
---------
Co-authored-by: saad <Saad>
* initial
* lint fix
* docs!
* fix CI
* ci and runner fix
* fix + docs!
* make CI pass
* country specific search
* remove country specific language
* fix UI
* address comments
* Address visa sponsor PR feedback
* Address remaining visa sponsor review feedback
* Harden visa sponsor provider validation
* feat(settings): add rxresume mode and v5 api key settings
* feat(server): add mode-aware rxresume adapter with auto v5-first selection
* refactor(server): route settings profile and pdf generation through rxresume adapter
* feat(api): support rxresume v4/v5 in onboarding and settings routes with ok/meta responses
* feat(client): add rxresume mode selector and v5 api key setup flow
* docs: document rxresume auto mode with v5-first self-hosted setup
* test: verify dual-mode rxresume support and ci parity checks
* comments
* services folder
* correct types for v5
* tests and docs fix
* Fix RxResume auto fallback and route API consistency
* warning for both being set
* simpler response
* onboarding component improvements, v5 check still not working
* fix list resume endpoint...
* fix api endpoints to latest v5 docs
* don't show the entire project field on v5
* remove auto entirely
* formatting
* ci green
* v5 has a different resume schema
* remove redundant check
* remove requirement that only one must be specified
* consolidate sections
* base resume can be v4 or v5
* saving now works
* status indicator
* actually render some pills
* reason for failure
* fix apikey verification
* dedupe isValidatingMode
* reefactoor
* simplification?
* refactor?
* ci passing
* remove auto from docs
* tailoring is schema dependent
* skills object tighter
* remove redundant text
* fix lint
* mode
* Deduplicate string cleanup helpers and not-found responses
* Enforce aliased imports for infra and shared modules
* Enforce @client/@server aliases for deep relative imports
* Deduplicate visa sponsor and location filter definitions
* Use shared city filter export in extractor location checks
* chore: move @types/canvas-confetti to devDependencies, remove unused get-tsconfig direct dep
* chore: configure knip with workspace entry points for all packages
* refactor(shared): split 1119-line types.ts into domain modules under types/
* refactor: remove llm-service.ts shim, migrate all import sites to llm/service directly
* refactor(settings): migrate 4 manually-resolved settings into conversion registry
* refactor: split gmail-sync.ts into gmail-api, email-router, and thin orchestrator
* refactor(orchestrator): extract useKeyboardShortcuts and usePipelineControls from OrchestratorPage
Splits the 840-line OrchestratorPage into a thin orchestration shell (~480 lines) by
extracting keyboard shortcut handling into useKeyboardShortcuts.ts and pipeline
control logic into usePipelineControls.ts. Net negative line count across all files.
* feat: create settings registry (Step 1)
Introduces a single source of truth for all settings, combining schema definitions, default logic, parsing, and serialization into a single configuration object.
* feat: derive schema, keys, and types from settings registry (Step 2)
Derives AppSettings nested shape, SettingKey DB union, and updateSettingsSchema Zod shape automatically from the settings registry.
* refactor: gut envSettings and remove settings-conversion (Step 3)
Replaces manual env arrays with registry-driven maps in envSettings.ts.
Deletes settings-conversion.ts since all parsing/defaults now live in the registry.
* refactor: simplify getEffectiveSettings with generic loop (Step 4)
Replaces ~334 lines of manual key-by-key unpacking with a generic registry-driven iteration loop (~40 lines). Models, typed, string, and virtual kinds are automatically derived.
* refactor: simplify settingsUpdateRegistry (Step 5)
Replaces ~350 lines of explicit per-key update handlers with a dynamic generic loop over the settings registry, properly routing persistence and side effects.
* refactor(settings): implement nested settings registry and clean up tests
- Migrate settings system to use a centralized nested registry (`settings-schema.ts`, `registry.ts`)
- Remove obsolete flat-to-nested conversion logic (`settings-conversion.ts`)
- Address Biome warnings by explicitly ignoring intentional `any` usage in generic runtime schema builder and registry logic
- Clean up unused variables in test files (`SettingsPage.test.tsx`) to achieve a 100% green CI pipeline
* refactor(settings): address PR comments on env data and registry parsing
- Narrow `getEnvSettingsData` return type to `Partial<AppSettings>` to satisfy strict typing and omit 'typed' registry entries
- Introduce `parseNonEmptyStringOrNull` for typed string settings so empty-string overrides cleanly fall back to defaults (matching original `||` logic)
- Add missing unit tests for registry parse/serialize helpers (JSON, bools, numeric clamping)
* feat(orchestrator): parallelize scoring and bulk rescore with bounded concurrency
* chore(docs): format versioned docs metadata for biome
* fix(orchestrator): address PR review feedback on async pool and bulk rescore
* ci(docs): run biome --write before docs version commit
